INDEX

Explanations

no or not conditions

The neuron fires on negation and prohibition cues (especially “no,” “not,” “don’t,” “may not,” etc.) in instructional or warning contexts.

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

Negative Logits

さっそく

-1.18

早速

-1.13

衬衫

-1.04

 wasn

-1.02

 hopefully

-0.98

落ち着いた

-0.96

 also

-0.96

 almost

-0.95

さてさて

-0.95

کنش

-0.94

POSITIVE LOGITS

nor

1.30

 this

1.19

 sondern

1.13

 karena

1.10

 هذا

1.09

ּוֹ

1.07

 ponieważ

1.05

or

1.05

adone

1.03

 cannot

1.02

Activations Density 0.046%